[PATCH 2/5] libxfs: refactor the btree size calculator code
Darrick J. Wong
darrick.wong at oracle.com
Mon Feb 1 13:14:53 CST 2016
On Mon, Feb 01, 2016 at 10:17:56AM -0500, Brian Foster wrote:
> On Fri, Jan 22, 2016 at 04:35:08PM -0800, Darrick J. Wong wrote:
> > Create a macro to generate btree height calculator functions.
> > This will be used (much) later when we get to the refcount
> > btree.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong at oracle.com>
> > ---
>
> The refactoring looks fine to me... though shouldn't this head into the
> kernel first and get backported to xfsprogs (or did I miss that
> somewhere)?
I think I've posted the kernelside version of this patch, though I suppose it
doesn't really become important until one starts adding the rmap/reflink stuff.
> That and one question below...
>
> > libxfs/xfs_bmap.c | 18 +-----------------
> > libxfs/xfs_bmap_btree.c | 9 +++++++++
> > libxfs/xfs_bmap_btree.h | 3 +++
> > libxfs/xfs_btree.c | 28 ++++++++++++++++++++++++++++
> > libxfs/xfs_btree.h | 3 +++
> > 5 files changed, 44 insertions(+), 17 deletions(-)
> >
> >
> > diff --git a/libxfs/xfs_bmap.c b/libxfs/xfs_bmap.c
> > index aef7cf3..c134765 100644
> > --- a/libxfs/xfs_bmap.c
> > +++ b/libxfs/xfs_bmap.c
> > @@ -172,25 +172,9 @@ xfs_bmap_worst_indlen(
> > xfs_inode_t *ip, /* incore inode pointer */
> > xfs_filblks_t len) /* delayed extent length */
> > {
> > - int level; /* btree level number */
> > - int maxrecs; /* maximum record count at this level */
> > - xfs_mount_t *mp; /* mount structure */
> > xfs_filblks_t rval; /* return value */
> >
> > - mp = ip->i_mount;
> > - maxrecs = mp->m_bmap_dmxr[0];
> > - for (level = 0, rval = 0;
> > - level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK);
> > - level++) {
> > - len += maxrecs - 1;
> > - do_div(len, maxrecs);
> > - rval += len;
> > - if (len == 1)
> > - return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) -
> > - level - 1;
> > - if (level == 0)
> > - maxrecs = mp->m_bmap_dmxr[1];
> > - }
> > + rval = xfs_bmbt_calc_size(ip->i_mount, len);
>
> We go from xfs_filblks_t to unsigned long here. Isn't the latter 4 bytes
> on 32-bit systems? Was that intentional?
Semi-intentional. The 'len' argument is the number of records you want to
store in the btree. For a bmbt, we're capped by di_nextents/di_anextents,
which limit us to 2^32 records and I think the methods in xfs_bmap.c won't
create an extent longer than MAXEXTLEN (2^21) blocks anyway.
OTOH I suppose the rmap btree can store more than 2^32 records when reflink
is turned on (not that you'd want to) so this could be widened to ULL.
--D
>
> Brian
>
> > return rval;
> > }
> >
> > diff --git a/libxfs/xfs_bmap_btree.c b/libxfs/xfs_bmap_btree.c
> > index 2ef1836..3c595e2 100644
> > --- a/libxfs/xfs_bmap_btree.c
> > +++ b/libxfs/xfs_bmap_btree.c
> > @@ -880,3 +880,12 @@ xfs_bmbt_change_owner(
> > xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
> > return error;
> > }
> > +
> > +xfs_filblks_t
> > +xfs_bmbt_calc_size(
> > + struct xfs_mount *mp,
> > + unsigned long len)
> > +{
> > + return xfs_btree_calc_size(mp, mp->m_bmap_dmxr,
> > + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK), len);
> > +}
> > diff --git a/libxfs/xfs_bmap_btree.h b/libxfs/xfs_bmap_btree.h
> > index 819a8a4..04bc6e2 100644
> > --- a/libxfs/xfs_bmap_btree.h
> > +++ b/libxfs/xfs_bmap_btree.h
> > @@ -140,4 +140,7 @@ extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,
> > extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
> > struct xfs_trans *, struct xfs_inode *, int);
> >
> > +extern xfs_filblks_t xfs_bmbt_calc_size(struct xfs_mount *mp,
> > + unsigned long len);
> > +
> > #endif /* __XFS_BMAP_BTREE_H__ */
> > diff --git a/libxfs/xfs_btree.c b/libxfs/xfs_btree.c
> > index 658f27e..d5f16c5 100644
> > --- a/libxfs/xfs_btree.c
> > +++ b/libxfs/xfs_btree.c
> > @@ -4081,6 +4081,34 @@ xfs_btree_change_owner(
> > return 0;
> > }
> >
> > +/*
> > + * xfs_btree_calc_size() -- Calculate the number of blocks needed to store
> > + * a given number of records.
> > + */
> > +xfs_filblks_t
> > +xfs_btree_calc_size(
> > + struct xfs_mount *mp,
> > + uint *limits,
> > + int maxlevels,
> > + unsigned long len)
> > +{
> > + int level;
> > + int maxrecs;
> > + xfs_filblks_t rval;
> > +
> > + maxrecs = limits[0];
> > + for (level = 0, rval = 0; level < maxlevels; level++) {
> > + len += maxrecs - 1;
> > + do_div(len, maxrecs);
> > + rval += len;
> > + if (len == 1)
> > + return rval + maxlevels - level - 1;
> > + if (level == 0)
> > + maxrecs = limits[1];
> > + }
> > + return rval;
> > +}
> > +
> > /**
> > * xfs_btree_sblock_v5hdr_verify() -- verify the v5 fields of a short-format
> > * btree block
> > diff --git a/libxfs/xfs_btree.h b/libxfs/xfs_btree.h
> > index 48cb251..b25ffd3 100644
> > --- a/libxfs/xfs_btree.h
> > +++ b/libxfs/xfs_btree.h
> > @@ -465,6 +465,9 @@ static inline int xfs_btree_get_level(struct xfs_btree_block *block)
> > #define XFS_BTREE_TRACE_ARGR(c, r)
> > #define XFS_BTREE_TRACE_CURSOR(c, t)
> >
> > +xfs_filblks_t xfs_btree_calc_size(struct xfs_mount *mp, uint *limits,
> > + int maxlevels, unsigned long len);
> > +
> > bool xfs_btree_sblock_v5hdr_verify(struct xfs_buf *bp);
> > bool xfs_btree_sblock_verify(struct xfs_buf *bp, unsigned int max_recs);
> >
> >
> > _______________________________________________
> > xfs mailing list
> > xfs at oss.sgi.com
> > http://oss.sgi.com/mailman/listinfo/xfs
>
> _______________________________________________
> xfs mailing list
> xfs at oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs
More information about the xfs
mailing list